From a Technical Standpoint, What Is Firework Flash Powder?

This article originally appeared in Fireworks Business, No. 246, 2004.

K. L. Kosanke and L. Weinman

perature attainable in pyrotechnic reactions, only Introduction solids and liquids incandesce, gases do not. As a In a recently published article on the regulatory practical matter, metal oxides are the only pyro- definitions of firework flash powder[1] it was con- technic reaction products that are not gaseous at cluded that none of those definitions provided suf- high temperature. Thus metal fuels need to be pre- ficient information to objectively establish wheth- sent in substantial quantity in flash powders. See er or not a pyrotechnic composition is a flash Table 1, which is a list of the boiling point of powder. That is to say, those definitions are all some of the pyrotechnic reaction products. The subjective to the extent that they depend on the table has somewhat arbitrarily been divided into intended use of the composition and none provide species with boiling points above and below a quantifiable measure that can be used to deter- 2500 °C. Note that the oxides of zirconium, mag- mine whether a particular pyrotechnic composi- nesium, aluminum and titanium top the list. tion is a flash powder. The purpose of the present article is to suggest a general approach that might Table 1. Boiling Point of Some Pyrotechnic be used as the basis for producing a quantitative [2] definition of flash powder. Reaction Products. The reason such an objectively quantifiable Reaction Product Boiling Point (°C) definition is needed is that – from both a regulato- ZrO2 ≈ 5000 ry and safety standpoint – flash powders are treat- MgO 3600 ed differently than other pyrotechnic composi- Al2O3 2980 tions. The rationale for this is that the hazards TiO2 2500–3000 posed by firework flash powders are generally SiO2 2230 significantly greater than most other commonly ZnO 1975 encountered pyrotechnic compositions. Accord- K2SO4 1689 ingly, both pyrotechnic manufacturers and regula- KCl 1500 tory enforcement personnel need to be able to un- K2CO3 d > 891(Tm) ambiguously identify whether a composition is or H2O 100 is not a flash powder. d = decomposes Background Discussion Tm = melting point

Before proposing a possible framework for a A second requirement for the production of the quantitative definition of firework flash powder, bright white flash of light characteristic of fire- consider the following non-quantitative definition work flash powders is a high reaction temperature. as a starting point for the discussion. To see why this is the case, consider Figure 1, Firework Flash Powder: Any active metal which presents information on the light produced [3] fueled pyrotechnic composition suitable for use in by incandescent bodies at various temperatures. a firework salute (or firecracker). Note particularly the two curves labeled 1727 C and 3727 C. While these two curves correspond to In considering this definition it is appropriate only doubling the absolute temperature, the inten- to consider why it requires that the pyrotechnic sity of light produced in the visible region in- composition be metal fueled. The bright white creases by a factor of approximately 500. (Be- flash of light characteristic of flash powders is cause the perception of light by humans follows produced by incandescence. At the reaction tem- an approximate logarithmic relation, the perceived

Selected Pyrotechnic Publications of K. L. and B. J. Kosanke Page 777 brightness produced by an incandescent object at Table 2. Heat of Reaction for the Burning of 3727 °C would be quite a bit greater, but it would Various Metal Fuels. [4] not be 500 times greater, than the light from the ° ° object at 1727 C.) Note also that at 1727 C al- Metal Reaction Heat of Reaction most all of the emitted light is in the long wave- Fuel Product (–kJ/mol) length (red) end of the visible range, whereas at Ti TiO2 945 3727 °C the wavelengths of light are more nearly Si SiO2 911 balanced across the visible range. The effect is Al Al2O3 838 that the perceived color of incandescent light pro- Mg MgO 602 duced at 1727 ºC will be distinctly orangish yel- Fe Fe O 412 ° 2 3 low, whereas that produced at 3727 C will be Zn ZnO 348 perceived as white. Accordingly, to produce a Pb PbO2 277 flash of light that is both bright and white, a high Cu CuO 157 reaction temperature is essential. Obviously metals do not all produce the same amount of thermal energy upon burning. While there is no universally accepted definition of an active metal, one possible definition could be those metals that burn with the production of the most abundant thermal energy. Accepting that definition for this article, those metals near the top of Table 2 would be the most active metals and would produce the highest reaction temperatures (i.e., those most capable of producing the brightest and whitest flashes of light). The second part of the above non-quantified definition for firework flash powder is that it be suitable for use in a firework salute. In essence this is just saying that the composition must be a reasonably violent explosive such that it is capable of producing the thunderous report (blast wave) expected for firework salutes when only Figure 1. Incandescent emissions at various tem- mildly confined in paper casings. To constitute a peratures. [3] reasonably violent explosive is saying in effect that: 1) the oxidizer must be reasonably effective, 2) the ratio of ingredients must be approximately The reaction temperatures for pyrotechnic correct, and 3) the particle size of the components compositions depend on a number of factors, but (most especially the active metal fuel) must be generally by far the most important is the thermo- sufficiently small. These are all important factors dynamic heat of reaction for the composition known to affect reaction rate.[5] Reasonably effec- (now more properly termed enthalpy of reaction). tive oxidizers include the chlorates, perchlorates To see why, for the production of high reaction and probably the nitrates commonly used in fire- temperatures, it is important that the metal fuel be work. The approximately correct ratio of ingredi- what is commonly described by a chemist as an ents would be those reasonably near the stoichio- active metal, consider the data in Table 2. Listed metric ratio of ingredients. Sufficiently small par- there are the heats of reaction for the burning of ticle size of ingredients is required such that most various metals (i.e., their combining with oxygen). of the fuel and oxidizer are fully reacted in the The table has somewhat arbitrarily been divided explosion and not blown clear such as in the form into metals with heats of reaction above and be- of sparks. low 500 kJ/mol. Note that titanium, silicon, aluminum and magnesium top the list.

Page 778 Selected Pyrotechnic Publications of K. L. and B. J. Kosanke Structure for a Potential Quantitative powder. One example of such a pyrotechnic com- Definition position is sparkler composition.[7] Thus some means of excluding such compositions from the At this point in the discussion, it seems appro- firework flash powder definition is needed. It is priate to begin the attempt to produce the prom- suggested that there be an additional requirement ised basis for a quantitative definition for firework in the definition of firework flash powder, one flash powder. This will be done in the form of an relating to the explosivity of the pyrotechnic com- actual definition; however, at this time the specific position. In the definition to be proposed, this ad- numbers to be used in the quantitative definition ditional requirement is stated as an exception. will not be included. (These values will need to be Thus the additional requirement would not always determined as a result of research and a consensus need to be considered. If a manufacturer is willing of opinion between representatives of the industry to accept that a pyrotechnic composition is a flash and enforcement agencies.) powder based on its metal content alone, that is Firework Flash Powder: Any pyrotechnic their choice and nothing more would need to be composition containing at least N1% of a metal considered. However, for any pyrotechnic compo- powder finer than N2 mesh. sition for which the limits on metal powder con- In terms of metal content, the above suggested tent were exceeded, yet the manufacturer believed definition is essentially that proposed to the US that composition was not sufficiently explosive so Bureau of Alcohol, Tobacco, Firearms and Explo- as to constitute it being classed as a flash powder, sives (ATF) by the American Pyrotechnic Associ- a relatively simple test could be conducted. That ation (APA), which follows: explosivity test would then determine whether that particular pyrotechnic composition was flash Flash Powder: Pyrotechnic compositions con- powder for the purposes of regulation. A possible sisting of one or more oxidizers such as potassium expanded definition for firework flash powder perchlorate, potassium chlorate, ammonium per- follows. chlorate, barium nitrate, or potassium nitrate combined with 25% or more by weight of metal Firework Flash Powder: Any pyrotechnic powder such as aluminum, magnesium, or magne- composition containing at least N1% of a metal sium/aluminum alloy (“magnalium”) or 30% or powder finer than N2 mesh, except when that py- more by weight of a combination of metal powder rotechnic composition does not produce an explo- combined with sulfur or antimony sulfide. The sion measuring at least N3 under specified test term “powder” means material capable of pass- conditions. [6] ing through a standard 275-mesh sieve. As with determining the values of N1 and N2, While the authors agree in large measure with the value of N3 and the standard explosivity test the APA approach, it is felt that test data needs to conditions will need to be arrived at by consensus. be produced before deciding on the exact percent- To reach a rational consensus, there will need to ages and particle size, whether nitrates should be testing of many compositions widely under- have different numbers, and what range of metals stood and accepted to be firework flash powder should be considered. Also, thought needs to be and many other compositions widely understood given to an additional requirement relating to the and accepted not to be firework flash powder. explosivity of the actual pyrotechnic composition Before completing this article, it is appropriate being considered. (This will be discussed further to include one possible example for the “specified below.) A definition such as that proposed by the test conditions” mentioned in the above definition. APA has the useful quality of allowing manufac- (Certainly a number of other tests might be con- turers to readily know, without the need for test- sidered as alternatives.) An apparatus much like a ing, whether or not any pyrotechnic composition version of a proximate audience concussion mor- they make is a firework flash powder and whether tar might be used, such as sketched in Figure 2. or not the regulatory and safety requirements for The means of ignition could be a short length of firework flash powder must be met. 3/32-inch (2.4-mm) Black Powder visco fuse (also The problem with the authors’ definition above called hobby, cannon or firework safety fuse). is that there are some pyrotechnic compositions This type of fuse is desirable because it is widely that are definitely not flash powders, yet they con- available and it provides a persistent ignition tain a high percentage of relatively fine metal stimulus, in contrast with electric matches. In Fig-

Selected Pyrotechnic Publications of K. L. and B. J. Kosanke Page 779 ure 2, the fuse hole is only very slightly larger than the fuse, such that the fuse will tend to stay in Sound place and relatively little of the test powder will Pressure enter the fuse hole. Also, it is suggested that the Meter fuse hole enter at the point where the larger drilled N hole – forming the test powder chamber – first 7 begins to taper to its point. This will allow the visco fuse to pass across the chamber, causing the N8 test powder to surround the fuse. The values for Test N4, N5 and the amount of powder used in the test Unit (N6) will need to be determined by consensus after testing. Figure 3. A sketch of the setup of the flash powder test unit and a sound pressure meter.

References 1) K. L. Kosanke and L. Weinman, “As Defined by Regulation, What Is Firework Flash Pow- der”, Fireworks Business, No. 244, 2004; also in Selected Pyrotechnic Publications of K .L. and B. J. Kosanke, Part 7 (2003 and 2004), Journal of Pyrotechnics, 2006.

2) CRC Handbook of Chemistry and Physics, Figure 2. A sketch of one possible approach to a th flash powder test unit. 75 ed., CRC Press, 1995. 3) K. A. Miller and W. E. Schneider, “Colorim- Any test composition that did not produce an etry: Methods and Tools”, The Photonics De- explosion would be considered not to be a flash sign and Applications Handbook, 1991. powder. Another possibility is to better quantify 4) B. Sturman, “An Introduction to Thermody- the output test such as illustrated in Figure 3. The namics”, Journal of Pyrotechnics, Issue 15, flash powder test unit would be placed on a flat 2002; also in Pyrotechnic Chemistry, Journal and relatively smooth hard surface, and the indica- of Pyrotechnics, 2004. tor of the power of the explosion could be a meas- 5) K. L. and B. J. Kosanke, “Factors Control- urement of the sound pressure level (dB) pro- ling Pyrotechnic Burn Rate”, Proceedings of duced. Here again, the distances N , N and sound 7 8 the Second International Symposium on pressure level (N ) will need to be determined by 9 Fireworks, 1994; also in Selected Pyrotech- consensus after testing. nic Publications of K.L. and B.J. Kosanke, Conclusion Part 3 (1993 and 1994), 1996. A quantitative definition of firework flash 6) J. Heckman, American Pyrotechnic Associa- powders is a complicated and involved subject. tion, private communication (2004). There are many issues that need to be considered 7) H. Ellern, Military and Civilian Pyrotechnic, and some are not strictly of a scientific nature. Chemical Publishing Co., 1968.

Some limited research and a third article are being considered that would allow this proposal to be taken to the next level by suggesting some possibly appropriate values for N1 through N9, as well as possibly considering other simple test methods.

Page 780 Selected Pyrotechnic Publications of K. L. and B. J. Kosanke